Statement Splitter

Dialect-aware SQL script splitting for multi-statement scripts, migration files, and procedural blocks.

Classes

class sqlspec.core.splitter.StatementSplitter

Bases: object

SQL script splitter with caching and dialect support.

__init__(dialect, strip_trailing_semicolon=False)

Initialize the statement splitter.

Parameters:

• dialect (DialectConfig) -- The SQL dialect configuration to use

• strip_trailing_semicolon (bool) -- Whether to remove trailing semicolons from statements

split(sql)

Split SQL script into individual statements.

Parameters:

sql (str) -- The SQL script to split

Return type:

list[str]

Returns:

List of individual SQL statements

class sqlspec.core.splitter.DialectConfig

Bases: object

Abstract base class for SQL dialect configurations.

__init__()

Initialize dialect configuration.

abstract property name: str

Name of the dialect.

abstract property block_starters: set[str]

Keywords that start a block.

abstract property block_enders: set[str]

Keywords that end a block.

abstract property statement_terminators: set[str]

Characters that terminate statements.

property batch_separators: set[str]

Keywords that separate batches.

property special_terminators: dict[str, Callable[[list[Token], int], bool]]

Special terminators that need custom handling.

property max_nesting_depth: int

Maximum allowed nesting depth for blocks.

get_all_token_patterns()

Get the complete ordered list of token patterns for this dialect.

Return type:

list[tuple[TokenType, str | Callable[[str, int, int, int], Token | None]]]

Returns:

List of tuples containing token types and their regex patterns

static is_real_block_ender(tokens, current_pos)

Check if END keyword represents an actual block terminator.

Parameters:

• tokens (list[Token]) -- List of all tokens

• current_pos (int) -- Position of END token

Return type:

bool

Returns:

True if END represents a block terminator, False otherwise

should_delay_semicolon_termination(tokens, current_pos)

Check if semicolon termination should be delayed.

Parameters:

• tokens (list[Token]) -- List of all tokens

• current_pos (int) -- Current position in token list

Return type:

bool

Returns:

True if termination should be delayed, False otherwise

class sqlspec.core.splitter.PostgreSQLDialectConfig

Bases: _EagerDialectConfig

Configuration for PostgreSQL dialect with dollar-quoted strings.

class sqlspec.core.splitter.OracleDialectConfig

Bases: _EagerDialectConfig

Configuration for Oracle PL/SQL dialect.

should_delay_semicolon_termination(tokens, current_pos)

Check if semicolon termination should be delayed for Oracle slash terminators.

Parameters:

• tokens (list[Token]) -- List of all tokens

• current_pos (int) -- Current position in token list

Return type:

bool

Returns:

True if termination should be delayed, False otherwise

static is_real_block_ender(tokens, current_pos)

Check if END keyword represents a block terminator in Oracle PL/SQL.

Parameters:

• tokens (list[Token]) -- List of all tokens

• current_pos (int) -- Position of END token

Return type:

bool

Returns:

True if END represents a block terminator, False otherwise

class sqlspec.core.splitter.TSQLDialectConfig

Bases: _EagerDialectConfig

Configuration for T-SQL (SQL Server) dialect.

Functions

sqlspec.core.splitter.split_sql_script(script, dialect=None, strip_trailing_terminator=False)

Split SQL script into individual statements.

Parameters:

• script (str) -- The SQL script to split

• dialect (str | None) -- The SQL dialect name

• strip_trailing_terminator (bool) -- If True, remove trailing terminators from statements

Return type:

list[str]

Returns:

List of individual SQL statements